Data Analysis

Perform the following analyses on data set: 1. Run the APPROPRIATE descriptive statistics for the variables Earn5yr, Test2005, Grad2005, SkillVal, STEM2010, FacSal05, FamIncL and FirstGen, and report them in a table. 2. How do the Median Earnings in 2011‐2012 of the graduates of the universities in your data compare with the Earning Power of Majors taught at each university? 3. Do universities with a higher percentage of students studying for STEM careers have higher Median Earnings? 4. For students entering in 2005‐2006, compare the average ACT/SAT math scores across the three groups of schools categorized by the proportion of first‐generation students (those whose parents did not go to college). Which differences are significant at the 5% level? What is the effect of accepting more first‐generation students on the average test score of schools? 5. Do universities with students from richer families (with higher annual incomes) enjoy higher Median Earnings? Suggest an explanation for your findings. 6. Is there a statistically significant relation between the proportion of first‐generation students attending a university and its Graduation Rate? Describe the relation in your own words, and suggest an explanation. 7. Are Faculty Salaries related to their students’ Median Earnings? Would paying the faculty more raise students’ Median Earnings? 8. Are the scores on standardized tests (SAT, ACT) related to family income? State one implication of this relationship. 9. Build a model to predict the median earnings in 2011‐2012 for the 2005‐2006 cohort (in 2014 USD). How well does your model fit the data? Identify the factors that have a significant effect on earnings. Explain, in one or two sentences for each factor, how the factor can affect earnings. SAVE the model‐predicted earnings of all the schools into a new variable (Save ‐ Unstandardized). 10. Compare the actual median earnings in 2011‐2012 with the model‐predicted earnings. Create a new variable called ValueAdd (actual ‐ predicted earnings) and report its descriptive statistics, including a histogram. Sort the schools in DESCENDING order of ValueAdd. Thus identify the top 5 value‐adding schools in your data set.